> reminds me of what we could have achieved if this ingenuity had been applied in another domain

It will; I think the reason we're seeing diffusion models applied to image generation first, is that it's a task that meshes well with the models. But also in general I think people will still be guided by the principle "use the right tool for the job" - this is just another tool. I doubt that the set of paths toward realization for any given needed creative imagery collapses to just "use a model"